Morse's Markov model of book use revisited

Over a decade ago, P. M. Morse suggested a simple Markov model of book use for a science library. Since then, his model has been modified and used in a number of applications, but the reliability of the basic assumptions underlying the model's parameters has not been tested. Utilizing 11 years of circulation transactions of a university library, it is shown that (1) Morse's model fits approximately 99% of the data for the whole collection and for three subject areas, and (2) contrary to his assumptions, one of the model's two parameters, a , is time dependent.